Label Refinery: Improving ImageNet Classification through Label Progression

TL;DR

This paper introduces Label Refinery, an iterative method to improve image classification accuracy by refining dataset labels, leading to significant performance gains across multiple models on ImageNet.

Contribution

The paper presents a novel Label Refinery process that iteratively refines labels, addressing issues like incompleteness and ambiguity, and demonstrates its effectiveness in boosting model accuracy.

Findings

Refined labels improve top-1 accuracy across various models.

Significant accuracy gains: AlexNet from 59.3% to 67.2%.

Other models also show notable improvements.

Abstract

Among the three main components (data, labels, and models) of any supervised learning system, data and models have been the main subjects of active research. However, studying labels and their properties has received very little attention. Current principles and paradigms of labeling impose several challenges to machine learning algorithms. Labels are often incomplete, ambiguous, and redundant. In this paper we study the effects of various properties of labels and introduce the Label Refinery: an iterative procedure that updates the ground truth labels after examining the entire dataset. We show significant gain using refined labels across a wide range of models. Using a Label Refinery improves the state-of-the-art top-1 accuracy of (1) AlexNet from 59.3 to 67.2, (2) MobileNet from 70.6 to 73.39, (3) MobileNet-0.25 from 50.6 to 55.59, (4) VGG19 from 72.7 to 75.46, and (5) Darknet19 from 72.9 to 74.47.